Top-down suppression of incompatible motor activations during response selection under conflict.

Top-down control is critical to select goal-directed actions in changeable environments, particularly when several options compete for selection. This control system is thought to involve a mechanism that suppresses activation of unwanted response representations. We tested this hypothesis, in humans, by measuring motor-evoked potentials (MEPs) elicited by transcranial magnetic stimulation (TMS) in a left finger muscle during motor preparation in an adapted Eriksen flanker task. Subjects reported, by a left or right button-press, the orientation of a left- or right-facing central arrow, flanked by two distractor arrows on each side. Central and peripheral arrows either pointed in the same (congruent trial) or in the opposite direction (incongruent trial). Top-down control was manipulated by changing the probability of congruent and incongruent trials in a given block. In the "mostly incongruent" (MI) blocks, 80% of trials were incongruent, producing a context in which subjects strongly anticipated that they would have to face conflict. In the "mostly congruent" (MC) blocks, 80% of trials were congruent and thus subjects barely anticipated conflict in that context. Thus, we assume that top-down control was stronger in the MI than in the MC condition. Accordingly, subjects displayed a lower error rate and shorter reaction times for the incongruent trials in the MI context than for similar trials in the MC context. More interestingly, we found that top-down control specifically reduced activation of the incompatible motor representation during response selection under high conflict. That is, when the central arrow specified a right hand response, left (non-selected) MEPs became smaller in the MI than in the MC condition, but only for incongruent trials, and this measure was positively correlated with performance. In contrast, MEPs elicited in the non-selected hand during congruent trials, or during all trials in which the left hand was selected, tended to increase more after the imperative signal in the MI than the MC condition. Another important observation was that, overall, MEPs were already strongly suppressed at the onset of the imperative signal and that this effect was particularly pronounced in the MI context. Hence, suppression of motor excitability seems to be a key component of conflict resolution.